LARGE-SCALE DISTRIBUTION OF GALAXIES: FILAMENTARY STRUCTURE AND VISUAL BIAS

We show that our eye is good at detecting global coherence and at the same time easily fooled into seeing large-scale filamentary features where none exist. For this reason an objective method of identification that also assesses statistical significance is necessary to confirm or reject the presence of filaments in the distribution of galaxies. We have found that a construction called the minimal spanning tree (MST), together with a data-permuting technique, works very well. We demonstrate the validity of this procedure on Glass patterns, a case where we know that our eyes are being fooled. It shows that we should not neglect visual cues but question their subjective interpretation. The MST procedure applied to the CfA redshift data finds statistical evidence for the presence of filamentary features.